We report the recent technical improvements in our microelectromechanical systems (MEMS)-based spectral-domain endoscopic OCT (SDEOCT) and applications for in vivo bladder imaging diagnosis. With the technical advances in MEMS mirror fabrication and endoscopic light coupling methods, the new SDEOCT system is able to visualize morphological details of the urinary bladder with high image fidelity close to bench-top OCT (e.g., axial/lateral resolutions, dynamic range) at a fourfold to eightfold improved frame rate. An in vivo animal study based on a porcine acute inflammation model following protamine sulfate instillation is performed to further evaluate the utility of SDEOCT system to delineate bladder morphology and inflammatory lesions as well as to detect subsurface blood flow. In addition, a preliminary clinical study is performed to identify the morphological features pertinent to bladder cancer diagnosis, including loss of boundary or image contrast between urothelium and the underlying layers, heterogeneous patterns in the cancerous urothelium, and margin between normal and bladder cancers. The results of a human study (91% sensitivity, 80% specificity) suggest that SDEOCT enables a high-resolution cross-sectional image of human bladder structures to detect transitional cell carcinomas (TCC); however, due to reduced imaging depth of SDEOCT in cancerous lesions, staging of bladder cancers may be limited to T1 to T2a (prior to muscle invasion).